Machine-learning prediction model for acute skin toxicity after breast radiation therapy using spectrophotometry

Prediction of acute skin toxicity in tomotherapy of breast cancer using skin DVH data

Investigation and quantification of the relationship between the skin dose-volume histogram (DVH) and the risk of acute skin toxicity in breast cancer patients undergoing Tomotherapy by regression modeling. This prospective study included 52 breast cancer patients treated with Tomotherapy in the dose range of 42.5-60 Gy to the planned target volume. Grading of acute skin toxicity in patients was assessed by the maximum score recorded in weekly follow-ups during and up to three months' post-radiation therapy using the Common Terminology Criteria for Adverse Events (CTCAE) v4.0 guidelines. A superficial layer with a thickness of 2 mm was designated as the Skin Representative Layer (SRL-2) on the Tomotherapy planning, and DVH was extracted for that. Then, multivariable and univariable logistic analyses were performed to identify the most predictive variables of acute skin toxicity from SRL-2 DVH values and patients' clinical parameters. The regression analysis identified V51Gy, representing the absolute SRL-2 volume receiving 51 Gy or more in physical dose, as the most predictive dosimetric parameter for grade 2-3 acute skin toxicity. The optimal cut-off value was 4.74 cc for the physical dose, with an Area Under the Receiver Operating Characteristic (ROC) Curve (AUC) value of 0.749, even when adjusted for clinical and treatment-related variables. The logistic model based on V51Gy demonstrated superior calibration, with a slope and R² value approaching 1, indicating better agreement between predicted and observed outcomes. The risk of acute skin toxicity during breast cancer Tomotherapy is correlated with the V51Gy parameter of skin DVH. Limiting V51Gy < 4.74 cc, or 23.7 cm2 of skin area, should keep the risk of grade 2-3 acute skin toxicity below 26%.

Practical recommendations for the management of radiodermatitis: on behalf of the ESTRO RTT committee

BACKGROUND

There is a substantial body of literature addressing the prevention, acute management, and follow-up care of radiation induced dermatitis (RID). The quality and application of this evidence, however, is inconsistent and its interpretation varies widely. While several national guidelines have been developed to standardise practices locally, many of these resources are not publicly available. On behalf of the European Society for Radiotherapy and Oncology (ESTRO) Radiation Therapist (RTT) Committee, an international writing group consisting of 12 experts from radiotherapy and two patient representatives composed a recommendation document for the management of RID.

MAIN BODY

The consensus for these recommendations was generated based on available international guidelines, and supplemented with evidence-based review articles on the topic. These recommendations focus on the prevention and practical management of early stage RID by avoiding skin trauma and maintaining hygiene. Addressing pain and inflammation in higher grades is also covered. The current literature refutes some of the traditional recommendations, especially restricting washing as well as the use of deodorant or the potential dose build-up of lotions which has been included and rectified in recent guidelines. In addition, the importance of grading the severity, including a baseline assessment is presented. The benefit of clear, and non-contradictory communication within the multidisciplinary team as well as patient involvement (e.g. PROMs or similar) is highlighted. Furthermore, the importance of recognising different skin types and skin tones, and the impact on how RID changes these in their appearance is stressed.

CONCLUSION

This document provides practical, actionable recommendations for the clinical management of RID, referencing the supporting literature. These recommendations have, however, identified a lack of high-level evidence, especially for agent-specific recommendations.

Development of Machine Learning Models for Predicting Radiation Dermatitis in Breast Cancer Patients Using Clinical Risk Factors, Patient-Reported Outcomes, and Serum Cytokine Biomarkers

BACKGROUND

Radiation dermatitis (RD) is a significant side effect of radiotherapy experienced by breast cancer patients. Severe symptoms include desquamation or ulceration of irradiated skin, which impacts quality of life and increases healthcare costs. Early identification of patients at risk for severe RD can facilitate preventive management and reduce severe symptoms. This study evaluated the utility of subjective and objective factors, such as patient-reported outcomes (PROs) and serum cytokines, for predicting RD in breast cancer patients. The performance of machine learning (ML) and logistic regression-based models were compared.

PATIENTS AND METHODS

Data from 147 breast cancer patients who underwent radiotherapy was analyzed to develop prognostic models. ML algorithms, including neural networks, random forest, XGBoost, and logistic regression, were employed to predict clinically significant Grade 2+ RD. Clinical factors, PROs, and cytokine biomarkers were incorporated into the risk models. Model performance was evaluated using nested cross-validation with separate loops for hyperparameter tuning and calculating performance metrics.

RESULTS

Feature selection identified 18 predictors of Grade 2+ RD including smoking, radiotherapy boost, reduced motivation, and the cytokines interleukin-4, interleukin-17, interleukin-1RA, interferon-gamma, and stromal cell-derived factor-1a. Incorporating these predictors, the XGBoost model achieved the highest performance with an area under the curve (AUC) of 0.780 (95% CI: 0.701-0.854). This was not significantly improved over the logistic regression model, which demonstrated an AUC of 0.714 (95% CI: 0.629-0.798).

CONCLUSION

Clinical risk factors, PROs, and serum cytokine levels provide complementary prognostic information for predicting severe RD in breast cancer patients undergoing radiotherapy. ML and logistic regression models demonstrated comparable performance for predicting clinically significant RD with AUC>0.70.

Training and temporally validating an NTCP model of acute toxicity after whole breast radiotherapy, including the impact of advanced delivery techniques

PURPOSE

The aim is to train and validate a multivariable Normal Tissue Complication Probability (NTCP) model predicting acute skin reactions in patients with breast cancer receiving adjuvant Radiotherapy (RT).

METHODS AND MATERIALS

We retrospectively reviewed 1570 single-institute patients with breast cancer treated with whole breast irradiation (40 Gy/15fr). The patients were divided into training (n = 878, treated with 3d-CRT, from 2009 to 2017) and validation cohorts (n = 692, treated from 2017 to 2021, including advanced RT techniques). In the validation cohort, patients were classified according to the delivery techniques into static (n = 404) and arc techniques (n = 288). Several clinical/technical information and DVHs of the "skin" (5 mm inner expansion from the body contour) were available. Skin toxicity was assessed during follow-up using the RTOG scale criteria. A multivariable logistic regression model was generated combining skin DVH and clinical parameters, using cross-validation methods that ensured high internal consistency and robustness. The performance of the model was tested in the validation cohort.

RESULTS

14.0 %/17.4 % of patients developed ≥ G2 toxicity, in the training/validation cohorts, respectively. The resulting multivariable logistic model included axillary lymph node dissection (OR = 1.58, 95 %CI = 1.01-2.48, p = 0.045), hypertension (OR = 1.54, 95 %CI = 1.04-2.27, p = 0.030) and skin V20Gy (OR = 1.008, 95 %CI = 1.004-1.013, p < 0.0001). The AUC of the model was 0.64/0.59 in training/validation, with better performance in the validation cohort if considering only V20Gy (0.62). The model showed satisfactory agreement between predicted and observed toxicity rates: in the validation group, the slope of the calibration plot was 0.96 (R2 = 0.6) with excellent goodness-of-fit (Hosmer-Lemeshow p-value = 0.99). Looking at each of the three predictors individually, only the role of V20Gy was confirmed in the validation group. Results were similar when considering patients treated with static or arc techniques.

CONCLUSION

An NTCP model for acute toxicity after moderately hypofractionated breast RT was trained. The model underwent temporal validation even for patients treated with advanced delivery techniques. Despite clinical differences and techniques, the confirmation of the dosimetry parameter in the validation cohort highlights its robustness and corroborates the hypothesis that skin DVH may assess the risk with the potential for improving plan optimisation.

Ultra-Hypofractionated Whole Breast Radiotherapy with Automated Hybrid-VMAT Technique: A Pilot Study on Safety, Skin Toxicity and Aesthetic Outcomes

Purpose

The most prevalent treatment-related side effect related to adjuvant radiotherapy (RT) for breast cancer is acute skin toxicity in the irradiated area. The purpose of this single-institution pilot study is to provide preliminary clinical results on the feasibility and safety of a breast ultra-hypofractionated radiation treatment delivered using an automated hybrid-VMAT technique. Skin damage was assessed both with clinical examination and objectively using a Cutometer equipment.

Patients and Methods

Patients received 26 Gy to the whole breast and 30 Gy to the tumoral bed in 5 fractions using an automated hybrid-VMAT approach with the option for the breath hold technique if necessary. Acute and late toxicities were clinically evaluated at baseline, 1- and 6-months after treatment using the CTC-AE v.5.0 scale. An instrumental evaluation of the skin elasticity was performed using a Cutometer® Dual MP580. Two parameters per patient, R0 (the total skin firmness) and Q1 (the elastic recovery), were registered at the different timelines.

Results

From June 2022 to January 2024, 30 patients, stage T1-T2, N0 were enrolled in the study. Four out of 30 (13.3%) patients reported G2 acute skin toxicities. At 6 months, G2 late toxicity was registered in 3 patients (10%). A total of 2160 measures of R0 and Q1 were recorded. At 1 month after treatment, no correlation was found between measured values of R0 and Q1 and clinical evaluation. At 6 months after treatment, clinical late toxicity ≥1 was strongly associated with decreased R0 and Q1 values ≥24% (p = 0.003) and ≥18% (p = 0.022), respectively.

Conclusion

Ultra-hypofractionated whole-breast radiotherapy, when supported by advanced treatment techniques, is both feasible and safe. No severe adverse effects were observed at any of the different timeframes. Acute and late skin toxicities were shown to be lower in contrast to data presented in the literature.

Metabolomics in Radiotherapy-Induced Early Adverse Skin Reactions of Breast Cancer Patients

Background

Early adverse skin reactions (EASRs) are common side effects of radiotherapy (RT) that impact the quality of life of breast cancer patients. This study used global metabolomics profiles of breast cancer populations to identify metabolic pathways and biomarkers significantly associated with RT-induced EASRs to identify potential targets for precision interventions.

Methods

We used a frequency-matched study design to identify pre-RT urine samples from 60 female breast cancer patients (30 with high and 30 with low EASRs) for metabolomic analysis by Metabolon Inc. using UPLC-MS/MS and GC-MS. Using MetaboAnalyst, we performed metabolomic data analysis and visualization on 84 candidate metabolites from 478 total compounds. We used the Oncology Nursing Society (ONS) Skin Toxicity Criteria (0-6) for EASRs assessment.

Results

Seven metabolic pathways were significantly associated with RT-induced EASRs, including alanine, aspartate, and glutamate metabolism (p = 0.0028), caffeine metabolism (p = 0.0360), pentose and glucuronate interconversions (p = 0.0028), glycine, serine, and threonine metabolism (p = 0.0360), beta-alanine metabolism (p = 0.0210), pantothenate and CoA biosynthesis (p = 0.0028), and glutathione metabolism (p = 0.0490). The alanine, aspartate, and glutamate metabolic pathway had the lowest false discovery rate (FDR)-adjusted p-value and the highest impact value of 0.60. Thirteen metabolite biomarkers were significantly associated with RT-induced EASRs.

Conclusion

Our data show that the alanine, aspartate, and glutamate metabolism pathways had the highest impact value on RT-induced EASRs. Future larger studies are warranted to validate our findings and facilitate targeted interventions for preventing or mitigating RT-induced EASRs, offering a promising direction for further research and clinical applications.

A dosiomics model for prediction of radiation-induced acute skin toxicity in breast cancer patients: machine learning-based study for a closed bore linac

BACKGROUND

Radiation induced acute skin toxicity (AST) is considered as a common side effect of breast radiation therapy. The goal of this study was to design dosiomics-based machine learning (ML) models for prediction of AST, to enable creating optimized treatment plans for high-risk individuals.

METHODS

Dosiomics features extracted using Pyradiomics tool (v3.0.1), along with treatment plan-derived dose volume histograms (DVHs), and patient-specific treatment-related (PTR) data of breast cancer patients were used for modeling. Clinical scoring was done using the Common Terminology Criteria for Adverse Events (CTCAE) V4.0 criteria for skin-specific symptoms. The 52 breast cancer patients were grouped into AST 2 + (CTCAE ≥ 2) and AST 2 - (CTCAE < 2) toxicity grades to facilitate AST modeling. They were randomly divided into training (70%) and testing (30%) cohorts. Multiple prediction models were assessed through multivariate analysis, incorporating different combinations of feature groups (dosiomics, DVH, and PTR) individually and collectively. In total, seven unique combinations, along with seven classification algorithms, were considered after feature selection. The performance of each model was evaluated on the test group using the area under the receiver operating characteristic curve (AUC) and f1-score. Accuracy, precision, and recall of each model were also studied. Statistical analysis involved features differences between AST 2 - and AST 2 + groups and cutoff value calculations.

RESULTS

Results showed that 44% of the patients developed AST 2 + after Tomotherapy. The dosiomics (DOS) model, developed using dosiomics features, exhibited a noteworthy improvement in AUC (up to 0.78), when spatial information is preserved in the dose distribution, compared to DVH features (up to 0.71). Furthermore, a baseline ML model created using only PTR features for comparison with DOS models showed the significance of dosiomics in early AST prediction. By employing the Extra Tree (ET) classifiers, the DOS + DVH + PTR model achieved a statistically significant improved performance in terms of AUC (0.83; 95% CI 0.71-0.90), accuracy (0.70), precision (0.74) and sensitivity (0.72) compared to other models.

CONCLUSIONS

This study confirmed the benefit of dosiomics-based ML in the prediction of AST. However, the combination of dosiomics, DVH, and PTR yields significant improvement in AST prediction. The results of this study provide the opportunity for timely interventions to prevent the occurrence of radiation induced AST.

Comparing Performances of Predictive Models of Toxicity after Radiotherapy for Breast Cancer Using Different Machine Learning Approaches

PURPOSE

Different ML models were compared to predict toxicity in RT on a large cohort (n = 1314).

METHODS

The endpoint was RTOG G2/G3 acute toxicity, resulting in 204/1314 patients with the event. The dataset, including 25 clinical, anatomical, and dosimetric features, was split into 984 for training and 330 for internal tests. The dataset was standardized; features with a high p-value at univariate LR and with Spearman ρ>0.8 were excluded; synthesized data of the minority were generated to compensate for class imbalance. Twelve ML methods were considered. Model optimization and sequential backward selection were run to choose the best models with a parsimonious feature number. Finally, feature importance was derived for every model.

RESULTS

The model's performance was compared on a training-test dataset over different metrics: the best performance model was LightGBM. Logistic regression with three variables (LR3) selected via bootstrapping showed performances similar to the best-performing models. The AUC of test data is slightly above 0.65 for the best models (highest value: 0.662 with LightGBM).

CONCLUSIONS

No model performed the best for all metrics: more complex ML models had better performances; however, models with just three features showed performances comparable to the best models using many (n = 13-19) features.

Mastectomy Skin Flap Stability Prediction Using Indocyanine Green Angiography: A Randomized Prospective Trial

BACKGROUND

The first step in successful breast reconstruction is obtaining a stable skin flap. Indocyanine green (ICG) angiography has recently been studied for its value and usefulness in predicting the stability of skin flaps; however, relevant prospective studies of its clinical efficacy are limited.

OBJECTIVES

The aim of this study was to prospectively investigate the clinical impact on breast reconstruction outcomes of the intraoperative use of ICG angiography.

METHODS

Between March and December 2021, 64 patients who underwent immediate breast reconstruction at the authors' institution were prospectively enrolled. They were classified into an experimental group (n = 39; undergoing ICG angiography) and a control group (n = 25; undergoing gross inspection alone). In the absence of viable skin, debridement was performed at the surgeon's discretion. Skin complications were categorized as skin necrosis (the transition of the skin flap to full-thickness necrosis) or skin erosion (a skin flap that did not deteriorate or become necrotic but lacked intactness).

RESULTS

The 2 groups were matched in terms of basic demographic characteristics and incision line necrosis ratio (P = .354). However, intraoperative debridement was significantly more frequent in the experimental group (51.3% vs 48.0%, P = .006). The authors additionally classified skin flap necrosis into partial- and full-thickness necrosis, with a higher predominance of partial-thickness necrosis in the experimental vs control group (82.8% vs 55.6%; P = .043).

CONCLUSIONS

Intraoperative ICG angiography does not directly minimize skin erosion or necrosis. However, compared to gross examination alone, it enables surgeons to perform a more active debridement during surgery, thereby contributing to a lower incidence of advanced skin necrosis. In breast reconstruction, ICG angiography may be useful for assessing the viability of the postmastectomy skin flap and could contribute to successful reconstruction.

LEVEL OF EVIDENCE

4.